Here’s why Twitter probably doesn’t mind that some of its users are robots

This post was updated on Tuesday at 3 pm ET after Quartz clarified its interpretation of the automated-account numbers that Twitter released in its 10-Q. We’ve asked Twitter for comment and will post it if and when we get any.

When Twitter released its second-quarter financial results, investors seemed pleased: revenue was up by more than 120 percent compared with the same period a year earlier, and user growth seemed favorable. But there was one number that was a bit of a red flag — namely, the number of users who never visit the website or use Twitter’s official apps, because they are pulling tweets in automatically through some kind of third-party service. The problem being that those users will never see any advertising, Twitter’s primary revenue source.

The company has now updated its numbers on automated accounts to note that they are lower than it previously said they were: A revised filing with the Securities and Exchange Commission states that while Twitter told investors that 14 percent of its estimated 270 million active users access the network via a third-party or automated service, further analysis shows that this number is closer to 11 percent, and only 8.5 percent use an automated method to read tweets.

Contrary to what Quartz originally reported, this doesn’t mean that more than 20 million of Twitter’s users are all bots, just that those accounts use some kind of external method for accessing tweets — a notification layer built into a phone, for example — as opposed to going to the Twitter website directly or using one of the company’s official mobile apps.

Not all automated accounts are bad

This means they aren’t necessarily the same as accounts that post tweets to the service automatically, in some cases based on an external signal of some kind — such as the @StealthMountain bot, which used to post a tweet whenever someone misspelled the phrase “sneak peek” (it is no longer active) or an account called @EarthquakesSF tweets whenever the USGS detects an earthquake. The conventional wisdom is that these kinds of accounts are useless because they can’t consume advertising. But some of them are likely very useful to Twitter.

For example, many brand accounts use automated services like SocialFlow to post tweets automatically when the algorithm detects that people are discussing a certain topic or product. That may seem like spam to some, but it also meets that advertiser’s needs of reaching out to potentially interested users — and if they see enough interest, they may even sign up for some promoted tweets.

The other potential benefit of bots is a bit more nebulous, and that is the engagement created by “Weird Twitter” bots such as @KimKierkegaard, which mashes up statements from celebrity tweeter Kim Kardashian with those from the existential Danish philosopher Soren Kierkegaard — or another popular account called @everyword, which did nothing but tweet a single word until it had gone through every word in the dictionary.

Again, these bots might seem like spam to some, but others (including me) enjoy the randomness they exhibit, and that likely keeps a certain number of users checking and refreshing Twitter. In the end, increasing engagement by actual humans is the number one challenge for Twitter, and if tweets from automated accounts help to do that, whether they are branded tweets or random bits of weirdness, then so much the better for Twitter and its investors.

Post and thumbnail images courtesy of Thinkstock / Ociacia