Twitter is being overwhelmed by spambots that undermine its value to advertisers—and to potential investors in its initial public offering—according to a University of Texas information scientist.
Some 24 percent of tweets are created by automated bots, according to one study, some of which are spam, some of which are not. Several news media headline feeds are automated (not @BW, though). A handful of bots were programmed to tweet about earthquakes in California. Still others provide useful and weirdly personalized services.
Still, the prevalence of nonpeople is a problem for the business’s revenue model, writes Andrew Whinston, director of the Center for Research in Electronic Commerce at the University of Texas at Austin. Whinston sent a one-page analysis that he co-authored with graduate student Shun-Yang Lee to Bloomberg Businessweek on Sept. 24. The short piece isn’t intended for an academic journal.
According to the New York Times, eMarketer estimates that Twitter will make $583 million in revenue from advertising in 2013 and $950 million in 2014. It makes additional revenue from selling the data about its raw tweet feed. But Whinston suggests that Twitter bots could be deployed to the same marketing ends.
We could imagine an army of bots that appear as real people tweeting about certain products. The sheer word of mouth effect generated from these tweets could easily beguile gullible Twitter users and achieve reasonable marketing goals—without ever paying Twitter. Therefore, advertisers could potentially seek help from programmers to write bots to trick Twitter’s authentication system and perform just as well as those costly promoted tweets and trending topics!
He suggests that Twitter protect its revenue model by tightening its spam filters and authentication mechanisms, either to eliminate bots or restrict them. Twitter is aware of the spambot problem but hasn’t been responding to journalists’ IPO-related questions, citing regulatory restrictions, since announcing its stock-sale plans in—of course—a tweet.