One in 12 Tweeps are bots
But only 5 per cent of accounts are spammers, says Twitter
In its latest SEC filing, Twitter has enumerated what its users have long believed: a small but significant slab of its accounts aren't humans.
According to the filing, Twitter says it's created a “new metric that is comprised of only such active users” that run updates “when there was no discernible user action involved”.
In the quarter ended June 30, up to around 8.5 per cent of users, the filing says, may be bots (note that it's not a definitive number, merely an estimate), and 11 per cent of all active users rely on third-party clients.
While users would probably react with “I know, right?” to the news, the metrics are of much greater import to advertisers, since bots are worthless targets for advertising.
As noted by Zac Seward over at Quartz, Twitter's estimate means that around 23 million of the claimed 271 million monthly active users (MAUs) are in some way automated.
That's not only spam accounts, since some of the MAUs might be automated news feeds (our own @theregister blends automated and real live human emissions). Regarding spambots, Twitter says: “We have performed an internal review of a sample of accounts and estimate that false or spam accounts represented less than 5 per cent of our MAUs” – without, however, revealing anything about the size of the sample or the methodology.
The financials show a net loss for the six months to June 2014 of US$277 million on revenues of nearly US$563 million. Revenue has more than doubled compared to the US$254 million for the same period last year, which delivered a net loss of US$69 million. ®