Twitter’s quickly growing (but still unprofitable) advertising business has gained lots of attention recently as the company prepares for an initial public offering. But selling promoted tweets is not the only way the social network generates revenue. Firms interested in gauging consumer sentiment about everything from computers to pop stars are shelling out millions of dollars to harvest and analyze the collective opinions of the Twitterverse.
In the first half of 2013, Twitter generated $32 million licensing its data—your tweets—to other companies, a 53% increase from the year before. Each month a small number of firms pay the social network for access to its “firehose,” the unending torrent of tweets from the service’s 215 million monthly active users. These companies, which Twitter calls data resellers, then turn around and sell access to portions of the firehose to social analytics companies like Klout and consumer brands such as Dell. Private tweets and direct messages are not included in the data deals.
“It’s all part of this movement to find out who the users are,” says Nora Barnes, director of the Center for Marketing Research at the University of Massachusetts Dartmouth. “We want to know who the dominant users are. We want to know what they’re saying. We want to know how to locate them. It’s the same no matter what business you’re in right now.”
Gnip, the first company to get access to Twitter’s data in 2010, now receives tens of thousands of dollars per month from large firms like IBM and Oracle, who then use the tweet databases in enterprise software for other companies. “The types of ways that businesses are using Twitter data has gone deeper and deeper,’ says Chris Moody, the CEO of Gnip. ”We’re seeing it in supply chain and inventory management. It’s not just consumer brands that are engaging on Twitter.”
Datasift, another company with access to the Twitter firehose, has also seen businesses use Twitter data in unexpected ways. One client, a sports equipment manufacturer, measured Twitter users’ chatter about various athletes to determine how many of each player’s jerseys to produce and sell. Another company, a music entertainment firm, tracked where the most people were tweeting about Justin Bieber to plan the stops on his tour in Turkey. “The early days of social are over and we’re finding that people are using social data in very sophisticated ways to make better decisions,” says Rob Bailey, Datasift’s CEO.
Neither Gnip nor Datasift would say how much they pay Twitter, but they are two of Twitter’s largest data partners. A Twitter spokesman declined to disclose a full list of the company’s data partners, though the company publishes a list of certified products that make use of Twitter data on its website.
Twitter’s data mining initiative has similarities with LinkedIn’s Recruiter program, which charges businesses as much as $8,000 per log-in to access the site’s full database of 238 million resumes. While LinkedIn manages Recruiter internally and generates massive revenue from it, Twitter has chosen a more hands-off approach that provides a small but straightforward revenue stream while leaving the major infrastructure costs to third parties.
As Twitter grows, it could decide to bring this service in-house. The company has spent the last two years buying up or killing off companies that use the Twitter API to create products that might compete with the company’s official website and mobile apps. Last year Twitter was sued by social analytics company PeopleBrowsr for cutting off the firm’s access to the data. In a settlement Twitter agreed to let PeopleBrowsr retain access until the end of 2013. But Gnip and Datasift say they have strong relationships with Twitter. In its IPO filing the social network said it expects data licensing to decrease from 13% of overall revenue currently to even less in the future.
Even if data becomes less important as a direct revenue source, there could be residual benefits from bolstering Twitter’s pedigree as a market research tool. “If I’m seeing Twitter data in my inventory management system and in my sales system…I’m going to begin to become a believer in the value of the platform,” Moody says. “I’m more likely to engage with my customers on the platform and I’m more likely to advertise as well.” Time will tell Moody’s theory is correct. Pre-IPO, some business executives are griping that Twitter doesn’t have the userbase to warrant large ad buys, according to the Wall Street Journal.
Still, other social networks are now making greater efforts to copy Twitter and turn user posts into searchable databases. Tumblr now licenses its data to both Gnip and Datasift. Facebook launched a new feature in September to provide news organizations with a firehose of publicly available posts. Datasift’s CEO says the company is being flooded with requests from smaller social networks wanting to distribute and monetize their data.
“It’s a huge area, and it’s one of the fastest growing areas because we can monetize it,” Barnes says. “When you’re in business today, information is gold. That’s currency.”