So that’s how Twitter makes money

Posted on March 13, 2012


Social networking site Twitter has sold two years worth of tweets to a marketing firm as it seeks new ways of monetising the site, a move which has enraged privacy campaigners.

The deal will allow businesses access to all tweets posted since January 2010, with UK firm DataSift becoming the first company to sell the data on to companies looking to target their marketing campaigns. Given that there are an estimated 250 million tweets a day, this is a sizeable amount of data.

The deal represents a step forward for Twitter, which has sold some data before, in finding ways to monetise the popular site with a future initial public offering (IPO) in mind given that it lags behind Facebook in terms of advertising revenue. The deal is positive news for companies looking to develop their marketing strategies, with the huge amount of data involved a goldmine for market researchers, yet has angered privacy campaigners at a time when the issue is gaining public attention.

In only six years Twitter has amassed more than 100 million monthly users, 300 million users worldwide and a growing influence on society and culture. In spite of this, however, Twitter as a business still has work to do to make itself ready for an IPO. Though they have tried to grow advertising revenues, it still lags behind Facebook, which is two years older. According to eMarketer Inc., Twitter’s advertising revenues last year were $139.5 million last year, 22 times smaller than those of Facebook which raked in $3.15 billion. Attempts to sell advertising revenues to Procter & Gamble (P&G), the biggest advertisers in the United States, failed last year, with companies appearing unsure as to how to use Twitter to enhance their brands.

“All [P&G] brands are asking questions about what to do with Twitter and how to leverage it; nobody really had a clear, lean answer,” said J.B. Kropp, the staffer charged with selling advertisements to P&G. This has remained a problem for Twitter as it seeks to make itself into a sustainable business worthy of its $8.4 billion valuation and look towards a future IPO. Twitter Chief Executive Dick Costolo has said that an IPO cannot take place until Twitter can demonstrate predictable earnings growth, which may not happen for a year or longer.

Previously, Twitter has been careful with selling advertisements for fear of alienating users, with advertisers paying around $120,000 a day to be amongst the site’s most popular topics. And though the company has changed policy in order to obtain more advertising revenue – targeting small businesses and looking to sell different types of advertisement, such as political ones – the move into the selling of data is an alternative means of gaining revenue. Though most tweeters believe that their tweets are unavailable after a week because they can no longer be searched on the site, Twitter in fact archives every tweet and has now agreed to sell all archived tweets from the last two years to DataSift. It is the first time that anybody has been able to access tweets from more than thirty days ago, and DataSift will pay Twitter a syndication fee each time it gives details of a tweet to a company.

“The news will surprise some,” said Graham Cluley of security firm Sophos. “Twitter has found another way to monetise its service, having partnered with a firm which will make it simple for market researchers working for big companies to search and analyse the last two years of your Twitter updates.”

DataSift has called its Twitter analysing cloud service Historics, and claim it allows companies to “improve forecasting of product-demand by analysing market sentiment towards their product compared to product-launches in the past.” DataSift’s marketing manager Tim Barker said the service would help businesses “understand their customers better”.

“We now live in an era where brands are shaped by what customers say and by accessing these tweets, companies can listen to their customers at scale,” he said. “It should come as no surprise to users that their tweets are archived – they can see every update they have ever sent on their timeline. Twitter was always created to be a public social network.”

The company charges up to £10,000 a month – it is £635 a month for an entry level package – to analyse tweets posted each day for any mentions about a client’s products or services. It claims to have almost 1,000 clients waiting to use its service for help with target advertising and developing marketing campaigns. According to DataSift CEO Rob Bailey, the waiting list for the service already includes 100 of the Fortune 500 companies. It will officially launch in April 2012.

“The sheer volume of data being produced by Twitter represents a huge challenge for companies trying to extract insights from past events,” said Bailey. “Historics solves this problem, providing businesses with a platform to intelligently filter and extract meaning from two years Twitter data.”

Yet it is just the first firm to be allowed access to Twitter’s “fire hose” – their full archives – at a cost, with others set to follow. Another service, Colorado-based Gnip, has been granted a licence this time around but only allows you to search thirty days of historical data. Other companies work by scraping data from Twitter, yet the new deal means that companies can pay to access the full data, improving the depth of their research and boosting Twitter’s revenues.

The announcement was not greeted with universal accord, however. Privacy campaigners, already hardly enamoured with tech giants after changes in Google’s privacy policy (see Special Report), reacted with dismay to the news. Online rights group Electronic Frontier Foundation called it “creepy”, while other groups criticised the transformation of Twitter from a social and cultural site to a marketing tool.

“Marketers will stop at nothing to get hold of your data,” said Justin Basini, chief executive of personal data and privacy rights company Allow. “This move shows that all those throwaway tweets have suddenly become a rich new revenue stream for Twitter, much in the same way that Facebook has monetised its offering. It has taken a stream of consciousness, analysed it, bottled it and sold it for a profit. And the worst thing is, you never knew it was going to happen.”

Nick Pickles, director of the Big Brother Watch campaign group, said that the deal turned Twitter users into products, while others felt that the deal turned Twitter from being a social network into a large market research project.

“People may consider tweets to be personal property but this deal makes clear they are not,” he said. “Our personal posts on social media are yet another way for advertisements to be better targeted and that’s a very lucrative industry. It’s clear that if you’re not paying for a service, you are not the customer – you’re the product.”

“People have used Twitter to communicate with friends and networks in the belief that their tweets will quickly disappear into the ether,” said Gus Hosein of Privacy International. “The fact that two years’ worth of tweets can now be mined for information and the resulting “insights” sold to businesses is a radical shift in the wrong direction. Twitter has turned a social network that was meant to promote global conversation into a vast market-research enterprise with unwilling, unpaid participants.”

Yet for all the complaints and talk of a new monetary direction for Twitter, this is not the first time that Twitter has offered access to its data for a fee. Google paid Twitter millions of dollars in order to index tweets for a real-time search experiment, a deal which expired last summer. Bing did the same. Many services have been built based on mining tweets for data trends. In 2010, Twitter agreed to share all of its tweets with the U.S. Library of Congress. Details of how this information is to be publically shared are still under development, though restrictions include a six-month delay and prohibition against using the information for commercial purposes. It seems that the deal with DataSift is both a way of increasing the revenue from the selling of data and outsourcing the act of selling to another company.

Posted in: News