The company has a grand scheme to monetize people who don’t know a retweet from a hashtag. This is Twitternomics.
It wasn’t a single meeting, an internal white paper or a blinding epiphany. Instead, it was a gradual realization — and a need for a life preserver— that led Twitter’s leadership to expand the definition of a “Twitter active user” beyond people who, um, use Twitter.
“I would say that in the middle of last year — maybe April or May — I started talking a lot internally about how we’ve got to do a better job of organizing content experiences for people who don’t log into Twitter, but see Twitter content all the time,” says CEO Dick Costolo.
Over the next few months, Costolo and crew crafted a theoretical construct that would increase Twitter’s audience without enlisting a single new user, envisioning a set of Dante-esque “concentric circles.” The bullseye of this target is the 284 million signed-in, monthly active Twitter users. In the circle around that we find people who come across Twitter but are not logged in — perhaps twice as many as in the inner circle. In another outside band are people who come across tweets embedded in blogs and news articles, as well as Promoted Tweets (Twitter ads), which may find their way onto properties beyond Twitter. “We think about everything we do in the context of this set of geometrically concentric circles,” said Costolo in an October 2014 earnings call.
Underlying that shift was a determination that the surprisingly successful business side of the company should not be held back by the company’s excruciatingly slow user growth. “There wasn’t a big metric that we looked at where it flipped one day and we’re like, aha! “ says Adam Bain, the company’s president of global revenue. “It was more a realization that all big monetization platforms are constantly looking to spread [outside their platforms]. Our point of view is that we have a superior mobile monetization unit and it could do more than even what we were doing inside of Twitter.”
Indeed, the Twitter-beyond-Twitter market is potentially huge — the company claims that in the third quarter of 2014, around 185 billion tweets were seen outside of the service (on other Web sites, print, television and even billboards). But can Twitter make money from all those tweets in the wild? Stock analysts like Richard Greenfield of BTIG Research, who last year published a blog item noting that “Off-Twitter” Monthly Active Users were growing ten times the rate of signed-in users, is skeptical. “I don’t understand how they monetize those Off-Twitter users!” he says.
But that’s exactly what Twitter plans to do. Just this week, for example, it announced the first syndication of Promoted Tweets, the ad system familiar to people who actually visit their Twitter timelines. Users of Flipboard and Yahoo Japan will be exposed to those ads, which will work just like the ones On-Twitter, except they might appear in the format of the new venue. (Flipboard already shows tweets in its own formalized style.)
The Off-Twitter strategy — which one might refer to as Stage Two of Twitternomics— is a classic instance of a “Got lemons? Make Lemonade!” gambit. Google has often been referred to as a “one-trick pony,” with almost all its revenues coming from its wildly successful search ads. Because Twitter’s successful ad product is limited by the service’s relatively low user base, the company needs to develop multiple tricks for its pony, which isn’t as big as Google’s to begin with. The would-be engines for this growth are three businesses being built to serve both an internal and extended Twittersphere.
In fall 2010 three hopeful entrepreneurs left Google to begin a company devoted to the placement of advertisements within smartphone apps. They gave a straightforward name to their mobile ad publishing company: MoPub.
It was a boom time for young companies with a sense of how to serve mobile ads. The original digital ad networks were still trying to figure out how to deal with the shift to mobile, and there was a clear opportunity for newcomers to disrupt the field, just as the DoubleClicks had marched in during the last revolution. Starting from a mobile-only perspective, MoPub and some others were able to tailor their businesses to the unique demands of the app world. Over a few years, MoPub developed an ad exchange where advertisers could target specific kinds of users and then bid in real time to insert ads into apps at just the right time. “We were one of the first companies that added a real-time bidding component to our system,” says John Egan, an early MoPub employee. “It allowed third-party buyers to say, Okay, this individual ad request is coming from somebody playing on an iPod Touch in the UK, and they are connected to wifi.” MoPub’s customers included WordPress, OpenTable and the game company Ngmoco.
Naturally, MoPub and the few other companies like it became prime takeover targets for big Internet and media companies who wanted that technology. (The MoPub founders themselves had left the Google division that had once been AdMob, a mobile ad company bought for $750 million in 2009.) So in one sense it was no surprise that in September 2013, Twitter bought MoPub for $350 million.
“You’ll notice that any company that has ‘ad’ before their name has had a successful outcome in acquisition,” says Janae McDonough, who is now senior director of exchange at Twitter, “exchange” being the company’s internal designation for MoPub. (McDonough herself had once worked for AdMeld, also bought by Google.) She might well have said the same thing for companies prefixed “Mo.”
Besides the $350 million, the MoPub people were excited because the Twitter purchase provided them the opportunity to make its targeting much more effective. As an independent company, MoPub was hampered by a lack of deep knowledge about an app’s users. It could target users of iPhones in geographic locations, for instance, but unless users shared personal data with their apps, MoPub couldn’t confidently deliver the microtargeting of a Facebook, a Google… or a Twitter.
She also notes that Twitter’s connections with big brands will expand MoPub’s reach — as well as the reach of Twitter advertising into the mobile world.
In other words, the MoPub acquisition allows the Twitter ad technology — so painstakingly created and so successful within the measly 284-million-user Twitter universe — to be used in the much broader world of anyone who uses ad-supported apps. This makes people who might never view a tweet into Off-Twitter users — not in a cultural sense, but a financial sense. (Last June, Twitter doubled down on the strategy by spending $100 million to buy TapCommerce, a mobile ad company specializing in technology that helps advertisers target specific groups of consumers.)
Apparently within Twitter there was a strain that thought the new acquisition should operate mainly in service of the company’s ad program. But Twitter now has determined that MoPub will also continue to operate as a quasi-independent ad network, even distributing ads from rivals such as Facebook. This means that on its own, the MoPub group should generate revenue. “There was a bit of an internal struggle, a lot of hard conversations that eventually caused us to go back to our original premise,” says Herman Yang, director of product management for the exchange group. “I think the interesting thing is that Twitter itself has moved closer to where MoPub was philosophically, to the idea that we need to embrace the entire ecosystem; we need to look beyond the Twitter app.”
MoPub is also in the center of Twitter’s plans to empower developers as they build and operate apps. Last October, at its own mobile developer conference, dubbed Flight, Twitter introduced a new Software Development Kit (SDK) that provides tools to help developers create mobile applications, whether they integrate Twitter functions into apps or not. (This approach puts Twitter in direct competition with its rival Facebook, which has a similar developer strategy.) The toolkit, dubbed Fabric, is of course geared for maximum integration of Twitter’s features into the app. The more that developers take advantage of this, the more “Off-Twitter” Twitter spreads.
When Twitter developed Fabric, it pointedly included MoPub as one of its pillars, a not-so-subtle motivation for developers to push Twitter ads into the apps they create. In effect, Twitter has drawn one more huge concentric circle, giving it a justification to regard just about anyone with a smartphone as a Twitter user. And the combination could be lucrative. One reason Twitter bought MoPub was to accelerate Off-Twitter growth of its ad platform. The integration with Fabric could make that happen more quickly and broadly. The lure of this plan is that it’s about monetizing the outer circles of Twitter, not the much smaller bulls-eye of signed-in users.
The CEO is pretty frank about this. “That was part of the reason why we brought in MoPub and TapCommerce, and some other assets,” Costolo says. “These are things that we’re starting to bring to ads not just on Twitter, but off Twitter.”
In essence, MoPub, Fabric and the other assets (including the data company Gnip and Twitter’s commerce initiative, which I’ll get to forthwith) constitute a massive hedge — a path to growth even if Twitter doesn’t grow its user base. “It’s one of the ways to make the business really resilient — as long as the universe of mobile applications leveraging our capabilities is growing, we can grow our business,” Costolo says. “If the mobile SDK is being distributed more broadly, and our ad exchange is part of that, it makes us even more resilient and anti-fragile. We go from, Gosh if this one thing doesn’t work, we’re in trouble, and make it more about, Gosh if this application ecosystem grows and we’re a core part of it, we’ll be able to grow the business.”
When Dick Costolo first arrived at Twitter as its COO in September 2009, one of the first things he did was complete some deals in progress to sell the company’s data (via the metaphorical “firehose” of all its activity) to big customers like Google and Microsoft. At the time, the thinking at Twitter was that the massive amounts of information the company was accumulating, as well as its proprietary ways of analyzing the corpus, might make up a significant chunk of the company’s revenues.
But as the ad business took off, Twitter’s data services faded as a potential contributor to its bottom line. So it was arguably a change in direction last April when Twitter announced that it was acquiring a 100-person Boulder, Colorado, social data provider called Gnip. (It’s “ping” spelled backwards.)
Gnip was already Twitter’s prime partner in the small but intense industry of delivering information from social networks to help corporations understand customers better — and sell more stuff and services to them. It was, in fact, a prime beneficiary of Twitter’s move in 2011 to cut off that aforementioned firehose to partners like Google. (Twitter had concluded that in those cases the accent on “frenemies” fell in the latter part of the word.) Founded in 2008 by two engineers, Gnip’s original mission was to provide a single API, or entrance point, to data in a wide variety of networks. It was referred to as “Grand Central Station for the Web.”
At the time of Gnip’s launch, MySpace was the biggest source of data. But eventually, Twitter became the dominant font of information. This was in part because its information was public, while important sources like Facebook sequestered a lot of its activity. But there was also that firehose — in 2011 Twitter gave Gnip favored access to that flow and Gnip was able to lock in customers as a result. “We were the first,” says Chris Moody, who became Gnip’s CEO that year. Twitter later opened the hose to a few other small companies. (Now, almost four years later, the news just came that Twitter is reportedly about to sell Google firehose access once again.)
Both parties were happy with that arrangement. Gnip kept building a wide customer base: “We have a distribution network touching over 95 percent of the Fortune 500,” says Moody. But everything changed in December 2013 when Apple paid $200 million to buy Topsy, another data analytics company with Twitter firehose access. Gnip’s customers worried that the company might be in play, jeopardizing their access to Twitter data. “I started getting phone calls from people like IBM and Adobe, saying, ‘Hey, Chris, what happens if you get bought?’” says Moody. “It doesn’t feel good when an executive is yelling at you and saying, ‘It’s really going to screw me over if you sell!’” Meanwhile, Twitter was worried that one of its large competitors might snap up Gnip and get access to its data.
So the two companies began talking — from Twitter’s point of view, perhaps grudgingly. “Day one, this was something Twitter did not want to think about,” says Moody. “You outsource stuff you don’t care about.” But over time, both organizations realized how important data was to the entire Twitter story.
So the obvious outcome occurred. Twitter later reported the cost as $134 million.
Much of the data Gnip delivered to companies was already affecting the way they used Twitter. Take the plight of T-Mobile in the period before it offered the iPhone. “They were bleeding customers because lots of people were leaving — but they didn’t know which ones they were,” says Moody. “So they took our data, mixed it with their own tools, and found out who was up for renewal. Twitter could tell who of those people was asking how the service is on Verizon, and who was talking about the iPhone.” T-Mobile could then offer those people better service deals to keep them.
Now that Gnip is part of Twitter, it can reach those people more easily. “So we’re helping T-Mobile make better decisions about who to target, and guess what?” says Moody. “They’re going to turn around and target them on Twitter!”
Going forward, Moody says that Gnip will not just be a catalyst for Promoted Tweets, but a revenue source of its own. Though Gnip never reported earnings as a private company, and Twitter does not break out its finances, Moody pointedly notes that prior to the acquisition, the company had needed only $6 million in funding, the last round coming in 2010. “That’s something that people in the venture world are floored by,” he says. “But the reality is it’s because we were generating meaningful revenue on our own to achieve these results.”
In fact, Moody believes that Gnip revenue will make a difference even in Twitter’s billion-dollar annual take. “We’re the fastest growing revenue piece of the business,” he says. “We absolutely think our revenue number will matter.”
A third potential source of new revenue for Twitter doesn’t involve going off-platform, but exploiting the commercial potential of the measly 284 million users who actually log in to Twitter — by selling stuff to them directly, via tweets.
The roots of Twitter commerce go back to 2010. After noticing the success of companies using the service to point people to deals (Dell Computer reaped a reported $65 million in sales through Twitter in 2008–2009), the company tried its own experiment, called Early Bird. Launched in July 2010, it offered @earlybird followers the chance to buy something at a discount, if they acted fast. By the end of September, 2010, Twitter decided that the @earlybird was too early, and smothered the project in its nest.
Over the next few years Twitter tried some isolated sales-by-Twitter with various sponsors like Starbucks or American Express, but nothing like a coherent initiative. But by early 2013, when the ad system was well underway, Costolo and Bain began thinking of other businesses Twitter could start, and relit upon the idea of using tweets for direct sales. Around that time Costolo was attending the Super Bowl in New Orleans, and had breakfast with Nathan Hubbard, the head of Ticketmaster.
Hubbard had taken that post in 2006 as a 30-year-old whiz kid; earlier in his life he’d been a singer-songwriter with five albums to his credit. Seven years into the job he was thinking of trying something else, and was asking Costolo if he was crazy to consider abdicating from such a plum perch. “No, you’re not,” said Costolo, who promptly began a discussion that would wind up with Hubbard’s joining the company several months later. It would be an understatement to note that Hubbard certainly was attracted to the product: “I felt that Twitter was at the intersection of the most important technologies and trends that are going to affect the rest of my life. It was the center of the universe from my perspective in that moment in time.”
So Hubbard joined Twitter to head e-commerce, going, he says, from overseeing three thousand employees to zero. He tapped an ex-Google engineer named Turan Jain to run product for him.
Twitter, Hubbard claims, is ground zero for “the three most important emerging trends in commerce.” The first is algorithmic curation. Twitter had already exploited what it calls its interest graph (the data that identifies people’s passions and curiosities) to allow its ad system to target ideal audiences. “Could we do that [for commerce] in a next-generation sort of way based on the amazing fingerprinting we can do?” asks Hubbard. “Could we target the right ad to the right user in the right moment [to sell them things]?”
The third trend is instant on-demand — the impulse that leads Amazon to contemplate drone delivery and venture capitalists to throw money at any startup claiming to be the “Uber of” just about anything. Hubbard thinks Twitter is perfect for impatient people. “It is the platform for real-time moments as they happen.”
Over the past year, Hubbard, whose commerce effort now involves around 100 Twitter employees, has built a system focusing on tweets with a “Buy Now” button similar to the ones on Amazon and iTunes. The semi-stealth nature of the project can be illustrated by the choice of the first tweet with the commerce button. Twitter’s ad platform started with a big-brand, culturally simpatico advertiser, Starbucks. But the first Buy Now button came in a tweet from Paramore, a rock band from Franklin, Tennessee. They were selling T-shirts.
Not long afterwards Twitter got a little bolder, arranging for controversial pundit Glenn Beck (something of a technophile) to sell books via the Buy Now button. Though it’s hard to say how much the campaign was influenced by his subsequent co-promotion on his Internet TV show, Beck apparently sold a ton of books. Hubbard is especially enthusiastic that almost all of those who bought the book on Twitter were newcomers to Beck’s extensive database of consumers of his wares. “We converted users from a social relationship to a transactional relationship, which is a big deal,” says Hubbard.
(On the other hand Twitter has no comment on the success or failure of Ashton Kutcher’s Buy Now button on a tweet to his 16 million followers — inviting them to purchase salt-and-pepper shakers. “We have had some real hits and we have had some real misses,” says Hubbard.)
Hubbard hopes that eventually Twitter — like Apple, with its iTunes — will as a matter of course be storing the credit cards for its hundreds of millions of users. But he doesn’t want to push too hard. “We’re wading very cautiously into this experiment,” he says. “Taking your payment details is a different kind of covenant with the user.”
(Has Hubbard considered exclusively using Square, a payment technology that just happens to be run by Twitter’s co-founder and board chairman Jack Dorsey? “No,” he says. “We want our users to pay however they want to pay. We want to put the right offer in front of the user in the right moment.” )
The key, says Hubbard, is not to be like Amazon — which is seen as a super-store that does the actual selling — but more like eBay, which is universally recognized as a lubricant to exchanges between buyers and sellers. “We’ve done a lot of work to make sure the user understands that the transaction process involves buying it from the merchant,” says Jain. “We don’t send the receipt — the receipt comes from the merchant.” Hubbard seconds this: “If they think it’s us [doing the selling] then the phone rings off the hook and people are in the lobby asking why they didn’t get their T-shirt in the mail.”
As the company gains experience, the experiments are getting more ambitious. During the post-Thanksgiving commerce period the Buy Now button popped up in timelines. Twitter also used the button on Giving Tuesday (following Black Friday), to help non-profits like the Nature Conservancy get instant donations.
“This is just the start,” says Hubbard. “Think of these as tools for merchants. The end goal is to allow our users to send an expression of demand into the Twittersphere and in real time have sellers of all types to be able to respond to you, based on all we know about you, where you are, what you like, the whole interest and follow graph.” Hubbard is also thinking of the possibility that besides appearing in tweets on one’s timeline, commerce tweets might also show up in a separate timeline of their own.
How significant will Twitter commerce be? Just ask Hubbard. “I think it’s going to change the way people search, discover and buy products,” he says. “We will be able to curate commerce experiences for you better than anyone in the world, and we will be the connective tissue between merchants and consumers on mobile phones. That’s how big I think it is.”
Forget about the concentric circles. If you believe Nathan Hubbard, the ultimate jewel of Twitternomics might lie in its sweet spot — the good old timeline.