Marketers on Twitter hope that their content gets people talking. They hope that the posts they publish will generate enough retweets and replies to persuade followers to discuss their products and talk about their brands. If a study released earlier this year by Pew Research Center is anything to go by, brand posts are indeed triggering responses. But no one is listening to them.
The study used NodeXL, an open-source template for graphing network data, to track Twitter conversations. By recording the connections between posts, retweets, followers and replies, the software is able to produce visual network maps of the relationships between users during conversations.
Altogether, the study mapped six different types of network — and turned up some interesting results.
Polarized crowds formaround contentious issues but they’re usually made up of two unconnected groups. Members of each group talk to each other, and each side links to different websites to support its position. When it comes to political discussion on Twitter, lots of people are talking but few people are hearing opposing arguments.
Tight crowds are made up of highly connected people, professionals who may know each other in real life and have strong interests in common. Their discussion often take place around conferences or hobby activities, and according to the report they “show how networked learning communities function and how sharing and mutual support can be facilitated by social media.”
Community clusters havemultiple centers of activity that may be triggered by the same event. The release of a new film in a series, for example, may inspire conversations among people looking for a something to watch at the weekend, while fans of the series will hold a very different kind of discussion among themselves on Twitter. Community clusters, the report argues, reveal a diversity of opinion and perspective on a social media topic.
Broadcast networks occur as people retweet the posts of news and media outlets. They act as loudspeakers that amplify the reach and power of mainstream publishers. Few members of a broadcast network will be connected to each other; they’ll be connected to the broadcast hub. Some members, though, have a strong interest in the topic and will form community clusters to discuss a tweet.
Support networks are hub-and-spoke structures created by corporate customer services teams on Twitter as they respond to complaints on the site.
The most interesting network that the study mapped however is brand clusters. These form around “well-known products or services or popular subjects like celebrities” as they’re discussed on Twitter. The maps drawn of these conversations show large numbers of “isolates,” participants whose posts show no connections with anyone one else.
“Well-known brands and other popular subjects can attract large fragmented Twitter populations who tweet about but not to each other,” the study says. “The larger the population talking about a brand the less likely it is that the participants are connected to one another. Brand-mentioning participants focus on a topic, but tend not to connect to each other.”
Alongside those isolates are often some community clusters. When Apple releases a new product, for example, Twitter will contain lots of posts from people declaring that they want one, and some discussions among fanboys talking up the new features. Cisco showed similar patterns: a large group of people mention the company on Twitter but they are largely disconnected from each other. Alongside those isolates are also some Cisco brand clusters — small, connected groups with few linkages to other groups, even those that also discuss Cisco. The content of the posts published in each group reference different URLs and use different hashtags.
Be A Hub And Build Bridges
For any company looking to promote its brand on Twitter, those maps represent a challenge. Part of the goal of a Twitter account will be to share information about the company but a large part will be to provoke the kinds of discussion that spread knowledge of the brand and deepen the relationship to it. If people are only posting that they want a new iPhone for Christmas and no one is bothering to reply, the reach and its effect will be limited.
The study did, however, point to one strategy that social media marketers could apply to turn those isolates into groups and those clusters into communities. The network maps that the study drew showed that key users occupied two kinds of vital strategic locations: they can be hubs, and they can be bridges.
Network maps locate the key people who are at the center of their conversational networks — they are “hubs” and they are notable because their followers often retweet or repeat what they say. Some people have links across group boundaries — these users are called “bridges.”
The brands themselves act as hubs. They’re creating the content that builds the clusters and around which the networks are formed. But a few people can act as bridges between clusters. As members of more than one grouping, they provide a way for messages to pass between different interested parties on Twitter and expand their reach.
Social media marketers need to be active hubs. They should be producing content that gets retweets and provokes discussion. But they should also be looking for the bridges that help to spread their brand messages across clusters, and making sure they know about their posts.
NodeXL may help to identify those bridges, and it is open source, but it’s likely that you know who those bridges are anyway. They’re key influencers, prolific tweeters with lots of followers who play an active role in discussions. Mention them in specific tweets to ensure that they’re aware of the content you want to broadcast. And you can go even further by mentioning two people active in different clusters in the same post. When they both followers see themselves mentioned in the same tweet there’s a chance that they’ll start talk to each other. And that’s a lot better than seeing your followers talking about your brand to themselves.