This post is about Twitter in specific, but also about social networks in general.
We had a nice Pi day yesterday at BOB and Patti's house, playing a few games, eating pie of various sorts and, as usual, sharing a few laughs. The subject of Twitter came up at one point, and I mentioned people who have hundreds of followers.
Chuck said something to the effect of "How can someone possibly follow that many people?"
It's something I've wondered myself, and we weren't even talking about the folks who have tens of thousands of follows. Truly, what sort of relationship does someone with that many follows (not followers, but folks they follow) have with their online friends? I want to address the effect of the number of followers on networks, efficiency, and value.
How should you feel when someone who follows 30,000 people follows you? How does that compare to someone following 3,000, 300 people, 30 people, or 3 people?
The person following 3 people is either your Twitter-resistant spouse, your mom, or a stalker.
That's a joke, but it makes sense because these are all people who are giving you a large portion of their attention, people with whom you have a close personal relationship.
The Twitter account with 30,000 is broadcasting, which is a one-way relationship, even if they're following you. It's easy to see why.
A Fraction of Attention
If you think of your time and attention as a limited commodity, then your partitioning of it to multiple people leaves less and less for each person. Normally, people can't easily see how much you value each individual connection of your network. However, with Twitter, it's right out in the open. In the account which follows 3,000 people, you are roughly 1/10th as valuable as if that person was following 300 people.
It's not an exact mathematics by any means, because there are tools that allow you to focus your attention on certain individuals and their may be some efficiencies in using Twitter which allow you to increase the overall value of your attention slightly with a large number of Twitter follows. However, attention is still limited and so a slight increase in your value as a member of the network is still divvied up over the entire network.
The fraction of their attention gets closer and closer to zero.
Fraction = Zero
And, coming back to filtering, let's say that a Twitter user is following 3,000 people but uses a tool like TweetDeck to single out 50 people whom have their actual attention. The rest are essentially filtered out; 2,950 may have a false illusion that they've got a fraction of this person's attention, when in reality they have almost none of their attention.
It seems like an efficient network should be working for you. The idea that you need to filter Twitter messages is an indication that your network may have failed you.
Picture a network with about 50 nodes, as I've depicted here. Assume each node is a person and assume that each connection is bidirectional: a pair of people that follow each other.
This graph is highly connected, but not saturated. In other words, despite all of the interconnections, many people are still not mutually following other people. I have artificially separated the graph into two groups of nodes, just to illustrate a contrast with the next graph.
This graph has two natural groups of nodes which are strongly interconnected, but between the two natural groups there are few interconnections. However, the two groups are still connected.
Which graph is more organized, and what does that mean for the quality of the communications in the network that the graph represents?
The second graph is more organized. I suggest that this causes natural filtering in the network. To illustrate, imagine this scenario: people are talking about their lunches, a bad joke, the weather, an illness, and then some event of national interest happens. In the first network, the smallest details are transmitted nearly universally over the entire network. When a more important event occurs, half of the network announces it causing everyone in the network to get a ton of repeat messages.
In the second graph, the less-universally-interesting messages only tend to stay within the tightly-connected sub-networks. How many people are going to repeat that Sally stubbed her toe? When a bigger message happens and is repeated, it will cross sub-networks. While it does so, the chatter within the chatter remains within sub-networks and only certain elevated items cross over.
Having many internally-connected sub-networks loosely connected with each other limits the noise and increases the signal because the community naturally filters out noise by not repeating it; the "signal" or anything worth repeating is more likely to cross to other sub-networks simply because it's getting repeated.
Instead of an individual having to filter out noise, the well-organized network does some of that work for you.
Spoilers are another good example. If you are close friends with a bunch of like-mined people when it comes to spoilers, you're unlikely to hear a spoiler from one of them. Not only are your "close friends" going to refrain from posting spoilers, spoilers are often spontaneously generated while people are watching a TV show; they're not the kind of thing that gets automatically repeated. So they're more likely to stay within a sub-network. If you have trouble avoiding spoilers, this might be a symptom of having too large a number of connections.
Adding By Subtracting
Earlier today, I went through and unfollowed some folks that I have been following. I based the decision partly on the value of their tweets and partly on whether they had huge follow lists.
If you follow me and 30 other people, you are likely having a conversation or some sort of relationship with me. If you follow 300 people but I occasionally get @replies from you, we're having a conversation. Conversations are of value.
If you follow and are followed by thousands of people, you're crossing over into broadcasting. If you're a broadcaster, I should more strictly judge the value of your messages because the conversation is missing; your effort in tweeting useful stuff must make up for that.
I removed people this morning for 2 reasons. Reason 1 was because their value had dropped. Reason 2 was so that I could add value to the other people I mutually follow. I added to my value to them by subtracting some of my follows.
I don't want to miss good stuff from the people I care enough about to follow, and I also want them to know that my limited attention is worth something.
Don't Be the Network
If you're following a ton of people to grow your network, consider the actual value of those connections. Being part of a network doesn't mean being connected to nearly everyone of value. People two connections away are still part of the network. Adding too many connections is an attempt to be at the center of the network, an attempt to be the network and take the place of the network. If you're busy being the network, you're not going to be of the same value.