Knowing a little network science can help you understand a lot about Twitter popularity.
Early moments
When Twitter released its new Moments feature earlier this year, I thought it would be an interesting exercise to discuss the development in my Internet Technologies class. So I asked my students about Moments during our daily ‘warm-up chat.’ I have found that this exercise gets them out of their shells, readies them for group-based engagement, and allows them a chance to vent about recent news. The response was amazing! We reached new heights of enlightenment by exploring how Twitter has solved the problem of getting important news to new users who don’t yet have the experience and knowledge required to select the perfect mix of accounts to follow. This skill is an art, much like that of an expert chef, and one I pride myself in having acquired through the years. Of course that is all a lie. All but two of my students unanimously felt underwhelmed: “I just don’t understand Twitter — 140 characters, what’s with that?” And even if they could get past the feelings of limitation, they still have another serious barrier to enjoying the microblogging platform.
The rich get richer effect
The founder of Scroll Kit, Cody Brown, in an article in TechCrunch, calls this barrier the ‘Root Injustice’ of Twitter — it’s hard for new users to get followers (Brown, 2014). If we pretend, for a moment, that users follow other users on Twitter at random, this effect is easy to visualize:
As a user gains more followers, they are more likely to gain followers-of-followers. This means that, once we factor in retweeting, a user with several followers is likely to have a much larger outreach than one with only a few, simply because having more followers means having more followers-of-followers to read their retweeted posts about their dreams and sudden interests in fusion cooking.
Friend of a friend
Cool. But users don’t just follow others at random. In fact, as Cody Brown explores in the TechCrunch article, Twitter even provides suggestions for users. If I am friends with your mom, your dad, and your uncle, then odds are I’ll end up friends with you too. So Twitter might list me on your homepage’s feed as someone you should really think about following too.
This human behavior (of having friends in common) leads to an increase in the effect we’ve been discussing: The more followers a user gets, the more likely they are to be listed as a suggestion. This nets them more followers, which in turn making them a suggestion for more people, and so on.
Further, it’s hard for a new user to get enough followers (without buying them) for this effect to really take off. At least Twitter ranks users in the suggestions list according to their engagement: a score based on interactions with their followers in the past month. So it should be hard to ‘fake’ your way to the top.
Unless, of course, you post really, really good clickbait (or have a robot do it for you).
I was shocked Clickbait, apart from being fairly senseless, can help us illustrate one emergent property of large, fast-growing systems: criticality. As Jennifer Ouelette explains in Wired and Per Bak explains in his book How Nature Works, sometimes something can happen that causes other things to happen
Imagine an electrical grid.
A failure at Point A can increase the ‘stress’ of the network nearby: Points B, C, and D. If we experience some bad luck that day, this new stress will cause B to fail as well, increasing the stress at C and D and maybe also at, say E, F, and G. I could keep this example going all the way through the alphabet. But if we let these failures keep propagating out, one by one, everyone will be sitting in the dark wondering what happened to their Netflix and chill (cause, you know, the Internet runs on electricity).
This cascade is called several things — percolation, propagation, self-organized criticality, virality. But the underlying principles are the same: When I do something, those near me are more likely to do that same thing.
A little goes a long way
This principle guides clickbait: Those ‘140 characters’ have been carefully crafted to elicit a retweet. This in turn causes the same 140 characters to appear on more feeds, causing more retweets, and so on, until tons of users in the Twitter-sphere see the same headline, “5 ways your cat could be messing up your life–number 4 will shock you!”
Although it’s easy to create a clickbait-sounding tweet, few Twitter users have the follower-base for a viral launch.
Those that do have plenty of followers. This means that the viral news that we see on social media can be heavily affected by those who are popular. But it can also be affected by large groups, say, during a revolution or two (Howard, 2015; Druzin, Lee, 2015). In either case, though, Twitter never really gave anyone any power or enabled them in any particular way, but instead offered them a chance to share information.
And if I’ve learned anything from studying how uncertainty affects decisions, it’s that computers change what we can do, but information changes what we will do.
Viral journalism
If what we publish has any effect on the world, then reaching more people gives us the power to amplify that effect. Sure, we, as publishers, would love to take advantage of networks that allow virality. But we as consumers should be cautious of the ability of others to take advantage of us through these same networks.
Shahid Mahmood said it best in 2011: “The world is not moved by the strength of the wealthy few but by the total sum of tiny pushes of hundreds of millions of people.” And if I can spend 140 characters getting 1.4 million people to make the tiny pushes I want, well, then I’m pretty much in control, aren’t I?
And this is percolation: popularity that begets popularity, tiny pushes that beget tiny pushes.
Bak, Per. How Nature Works: the science of self-organized criticality (Copernicus, 1999).
Howard, Phil. “Project on Information Technology and Political Islam.” Retrieved from Philhoward.org (2015).
Druzin, Bryan and Li, Jessica. “The Power of the Keystroke: Is Social Media the Radical Democratizing Force We’ve Been Led to Believe It Is?,” Harvard Human Rights Journal, 1, Vol. 28 (2015).
Knowing a little network science can help you understand a lot about Twitter popularity.
Early moments
When Twitter released its new Moments feature earlier this year, I thought it would be an interesting exercise to discuss the development in my Internet Technologies class. So I asked my students about Moments during our daily ‘warm-up chat.’ I have found that this exercise gets them out of their shells, readies them for group-based engagement, and allows them a chance to vent about recent news. The response was amazing! We reached new heights of enlightenment by exploring how Twitter has solved the problem of getting important news to new users who don’t yet have the experience and knowledge required to select the perfect mix of accounts to follow. This skill is an art, much like that of an expert chef, and one I pride myself in having acquired through the years. Of course that is all a lie. All but two of my students unanimously felt underwhelmed: “I just don’t understand Twitter — 140 characters, what’s with that?” And even if they could get past the feelings of limitation, they still have another serious barrier to enjoying the microblogging platform.
The rich get richer effect
The founder of Scroll Kit, Cody Brown, in an article in TechCrunch, calls this barrier the ‘Root Injustice’ of Twitter — it’s hard for new users to get followers (Brown, 2014). If we pretend, for a moment, that users follow other users on Twitter at random, this effect is easy to visualize:
As a user gains more followers, they are more likely to gain followers-of-followers. This means that, once we factor in retweeting, a user with several followers is likely to have a much larger outreach than one with only a few, simply because having more followers means having more followers-of-followers to read their retweeted posts about their dreams and sudden interests in fusion cooking.
Friend of a friend
Cool. But users don’t just follow others at random. In fact, as Cody Brown explores in the TechCrunch article, Twitter even provides suggestions for users. If I am friends with your mom, your dad, and your uncle, then odds are I’ll end up friends with you too. So Twitter might list me on your homepage’s feed as someone you should really think about following too.
This human behavior (of having friends in common) leads to an increase in the effect we’ve been discussing: The more followers a user gets, the more likely they are to be listed as a suggestion. This nets them more followers, which in turn making them a suggestion for more people, and so on.
Further, it’s hard for a new user to get enough followers (without buying them) for this effect to really take off. At least Twitter ranks users in the suggestions list according to their engagement: a score based on interactions with their followers in the past month. So it should be hard to ‘fake’ your way to the top.
Unless, of course, you post really, really good clickbait (or have a robot do it for you).
I was shocked Clickbait, apart from being fairly senseless, can help us illustrate one emergent property of large, fast-growing systems: criticality. As Jennifer Ouelette explains in Wired and Per Bak explains in his book How Nature Works, sometimes something can happen that causes other things to happen
Imagine an electrical grid.
A failure at Point A can increase the ‘stress’ of the network nearby: Points B, C, and D. If we experience some bad luck that day, this new stress will cause B to fail as well, increasing the stress at C and D and maybe also at, say E, F, and G. I could keep this example going all the way through the alphabet. But if we let these failures keep propagating out, one by one, everyone will be sitting in the dark wondering what happened to their Netflix and chill (cause, you know, the Internet runs on electricity).
This cascade is called several things — percolation, propagation, self-organized criticality, virality. But the underlying principles are the same: When I do something, those near me are more likely to do that same thing.
A little goes a long way
This principle guides clickbait: Those ‘140 characters’ have been carefully crafted to elicit a retweet. This in turn causes the same 140 characters to appear on more feeds, causing more retweets, and so on, until tons of users in the Twitter-sphere see the same headline, “5 ways your cat could be messing up your life–number 4 will shock you!”
Although it’s easy to create a clickbait-sounding tweet, few Twitter users have the follower-base for a viral launch.
Those that do have plenty of followers. This means that the viral news that we see on social media can be heavily affected by those who are popular. But it can also be affected by large groups, say, during a revolution or two (Howard, 2015; Druzin, Lee, 2015). In either case, though, Twitter never really gave anyone any power or enabled them in any particular way, but instead offered them a chance to share information.
And if I’ve learned anything from studying how uncertainty affects decisions, it’s that computers change what we can do, but information changes what we will do.
Viral journalism
If what we publish has any effect on the world, then reaching more people gives us the power to amplify that effect. Sure, we, as publishers, would love to take advantage of networks that allow virality. But we as consumers should be cautious of the ability of others to take advantage of us through these same networks.
Shahid Mahmood said it best in 2011: “The world is not moved by the strength of the wealthy few but by the total sum of tiny pushes of hundreds of millions of people.” And if I can spend 140 characters getting 1.4 million people to make the tiny pushes I want, well, then I’m pretty much in control, aren’t I?
And this is percolation: popularity that begets popularity, tiny pushes that beget tiny pushes.
Bak, Per. How Nature Works: the science of self-organized criticality (Copernicus, 1999).
Howard, Phil. “Project on Information Technology and Political Islam.” Retrieved from Philhoward.org (2015).
Druzin, Bryan and Li, Jessica. “The Power of the Keystroke: Is Social Media the Radical Democratizing Force We’ve Been Led to Believe It Is?,” Harvard Human Rights Journal, 1, Vol. 28 (2015).
