One Big Doubt Hanging Over Twitter’s IPO: Fake Accounts; With Robot Accounts Spitting Tweets, Marketers May Want More Certainty They’re Reaching Humans

Updated October 3, 2013, 7:40 p.m. ET

One Big Doubt Hanging Over Twitter’s IPO: Fake Accounts

With Robot Accounts Spitting Tweets, Marketers May Want More Certainty They’re Reaching Humans

TOM GARA

At 4:45 p.m. on Thursday, the Twitter account for Mashable—one of the earliest movers into the now endless world of social media news sites—sent out its 60th tweet for the day. The tweet itself wasn’t particularly interesting, But what happened next was a small window into one of the biggest challenges Twitter will face as it seeks to convince investors that its more than 215 million users are one of the web’s most lucrative—and undeveloped—advertising audiences.Within minutes, hundreds of other Twitter accounts reposted the Mashable tweet word for word—more than 400 of them within the first 15 minutes. Scrolling through the never-ending list of those copycats is like a deep dive into the Twitter netherworld, one that regular users would seldom find themselves in.

Beyond the Biebers and Mashables, the CNNs and Kardashians, is a Twitter world populated by millions of accounts of questionable legitimacy.

They range from entirely robotic (and often incomprehensible) spammers to more cleverly programmed accounts spitting out tweets designed to find their way into the occasional search results or discussion thread.

One reason these bots exist is the vibrant gray market in buying and selling Twitter accounts that have a following. Just like lists of working email addresses can be sold to junk mail marketers, so too can Twitter accounts with thousands of followers.

Another reason so many of these accounts are out there is that they are so easy to create and automate, thanks to the way Twitter allows automatic posting using third-party applications.

Twitter Inc. has long said it is working to stamp out spam on the site, and in April 2012 announced it had filed a lawsuit in federal court in San Francisco against a group of companies it said were the most aggressive spammers of the site.

“While spam is a small fraction of the incredible content you can find on Twitter, we know just how distracting it can be,” the company said at the time.

A Twitter spokesperson declined to comment on Thursday. In the IPO documents filed Thursday, Twitter estimates that “false or spam accounts” make up less than 5% of the site’s 215 million monthly active users.

With that said, its definition of active user varies from the typical “unique” user measured by most websites.

“We treat multiple accounts held by a single person or organization as multiple users for purposes of calculating our active users,” the company said. The company also said automated accounts that send regular tweets could be counted as active users.

Such accounts would fit into a broad category of Twitter usage that is neither human nor junk: Automated accounts set up by companies, publishers, charities or just plain Twitter hobbyists.

They believe they are creating useful services by tweeting daily weather forecasts for any given city, or show times for local movies, or all the news headlines related to a certain topic.

Combined, they create a strange situation: Twitter is essentially an advertising-supported media company, but a segment of its user accounts can’t be considered “eyeballs” in the traditional sense of the term.

Working out where the humans end and the robots begin—and separating the good robots from the bad—will be a big question facing the post-IPO Twitter.

While a site like Facebook has a fairly well-defined way to categorize its users—there are personal profiles for individuals, brand pages for businesses, fan pages for celebrities—the boundaries in Twitter are less clearly delineated.

Twitter has long spoken of a philosophical difference with Facebook in its approach to identity, with Twitter believing it is in the best interest of users and the service as a whole to allow multiple accounts per user, as well as pseudonymous and parody accounts. Facebook has been more insistent on genuine identities, and argues this is one of its major advantages as an advertising medium.

That difference in philosophy leads to some of the great joys of Twitter. There are parody accounts, posting in character as the Big Ben clock tower or as Kim Kierkegaard, a hybrid of Kim Kardashian and the Danish existentialist Soren Kierkegaard. There are conceptual accounts, tweeting World War II in real time or working their way through every word in the English language, one at a time.

But at the other end of the spectrum, Twitter’s looser concept of identity also means the site is populated with plenty of meaningless junk.

Those accounts are adding little but inflation of closely watched metrics, like the number of tweets posted on a certain topic, the number of followers on any given account, or the number of retweets of a post.

Consider @Web_Consult_, one of those reposting Thursday afternoon’s Mashable tweet. It has 433 followers, follows 571 people, and has posted more than 34,000 times since it began in September 2011—about 50 times a day.

Its most recent 20 tweets were all nothing but Mashable headlines. Scroll quickly through its most recent few hundred and it looks like that is all it does.

There are a couple of plausible explanations for what @Web_Consult_ is up to. It could be a regular person, trying their best to use Twitter and build up a following. The other is that this account is a robot.

If the former is true, then anybody analyzing the quality of the Twitter audience will need to factor in that some users are engaging with the site in a pretty limited way. Attempts to contact @Web_Consult_ were unsuccessful. It was still posting Mashable headlines as of Thursday evening.

Either way, for brands keeping an eye on their follower numbers or tracking the volume of discussion around a given topic, is there much value in counting an account like @Web_Consult_ alongside regular users?

Twitter is likely already being asked these kinds of questions by brands focusing more of their promotional efforts on the service. And as the company begins to take money from the public, and answer to that public, the big question is: What more can the company do about these kinds of accounts.

Facebook grappled with a similar challenge, and as a public company, now discloses what it believes to be its number of fake or duplicate accounts (as of the end of 2012, its guess was 76 million, or about 7% of its total membership base).

Twitter is now disclosing a similar estimate of false users, but has an even more important question to answer:

If its hundreds of millions of users range from everyday people to a constellation of robots and spammers, how can advertisers and the rest of its users work out who’s who?