When Twitter did an analysis of bot behavior and made the claim that less than 5% of all accounts are bots, did they factor only active accounts in that equation or did they include even inactive accounts or accounts that just log in and read tweets?
Does anyone know?
Conversation
Replying to
Thanks (and thanks for your awesome work). I think it makes a big difference on how the end number is derived. I'm really curious.
1
4
Show replies
Replying to
In the past there was a bit more nuance, easier to understand because it was more clearly defined but people misinterpreted it: time.com/3103867/twitte
As far as I can tell, now it's an estimate based on a sample (unclear how) from the quarterly report: investor.twitterinc.com/financial-info
read image description
ALT
1
2
2
In 2014 "approximately 8.5% of all active users used third party applications that may have automatically contacted our servers for regular updates without any discernible additional user-initiated action." makes perfect sense, "people used automation that worked in background"
1
1
2
Show replies
Replying to
And do they mean only "bad" bots or include "good" ones like weather, traffic, police apps etc.?
Replying to
They include accounts that can be reached with ads per month. So no inactive accounts. This metric cannot be independently audited the way Elon would like to, because there is no way to know who sees an ad. Twitter could for example hide ads from suspected spammers.
Show additional replies, including those that may contain offensive content
Show