In recent months, I’ve been asked to talk about the theory and practice of creating realistic Twitter bots, and it’s an area we know well.
The vast majority of bots you will come across on Twitter are the obvious type:
- Re-spun RSS feeds
- A link in every tweet
- Reaching out with no “conversation”
- Direct messages en masse
Really, they are the equivalent of e-mail spam or a bad mail merge letter.
Some people however have started to take the practice into more natural profile creations and these are far harder to spot. As long as they aren’t causing a problem, most people will be unaware of the subtle messages they give out and the data they collect. But if you don’t like the idea of anything automated interacting with you, then here is my guide to spotting the tell-tale signs.
Posting via the API
If they always post via the Twitter API, they are most likely a bot, especially if the API tweets are their normal day-to-day activities plus links.
They Follow Just Under 2,000 people (+/- 10 Percent)
This is a throttle point imposed by Twitter to stop the creation of automated accounts with many thousands of followers. Don’t rely on this though, as once they get more than 2,000 there aren’t any more realistic throttle points for them to follow.
They Don’t Engage in “Conversation”
Bots often await an engagement (a reply) before they signal a human to step in and engage directly. This often leads to a delay that doesn’t look natural.
Twitter Is Their Only Network
Normal people usually have accounts on other networks: Facebook, LinkedIn, etc. Bots never have a real life away from their jobs as a bot.
They Post When They Should Be Asleep
If they have stated they’re in the U.K. yet they post tweets at 5 a.m. U.K. time, it’s a good bet they aren’t sleep-tweeting. It’s more likely that they’re a humming air-conditioned server somewhere dark and peaceful.
Less Than 10 Percent of Their Tweets Are Me, Me, Me
Most people talk about their day on Twitter. And if you’re like me, you tell the world a hell of a lot more than you really should. Bots don’t often say what they had for tea, they rarely have an opinion on a film, they don’t drink, and above all else they don’t care about you and your tweets.
If you know what’s possible and you review who follows you and who you follow in return, it’s easy to see those accounts that are either automated or are only using the network for promotion. If that offends you, remove them. If it’s nasty, pointless spam, report them.
Luckily, it’s rare to see bots that talk about their day, reach out to you and engage, reply to conversations, and have a life outside Twitter. If someone is creating such things, they would be very hard to spot in the noise of social media, not that I have experience of that you understand…
Meet Paul Madden at SES San Francisco during Connected Marketing Week. He will be participating in the panel, “Twitternation & Automation,” on Wednesday, Aug. 18, 2010.
Time is running out to feature your company in our inaugural Mobile Vendor Reader Survey.
Marketers create personas to better understand their target audience and what it looks like. If marketers can understand potential buyer behaviors, and where they spend their time online, then content can be targeted more effectively.
What’s behind a successful data-driven marketing strategy?
Audience targeting can be challenging in social media, especially when brands make quick assumptions about their target users. How can you avoid generalisation and what are the real benefits of it?