The 6 Types of Twitter Conversations

Article Featured Image

Information scientists at the Social Media Research Foundation (SMRF) have long been tracking the types of conversations taking place across social media. In five years, SMRF has created roughly 1.5 million conversation maps. With a sample that massive, patterns were "bound to emerge," says SMRF cofounder Marc Smith.

Using all of this information, SMRF has identified six categories of conversations that take place on social media sites. Smith admits that there could be more, but that research is still ongoing.

The following conversation categories have been advanced so far:

  • polarized crowds, consisting of users who discuss polarizing topics, rely on sources of information, and typically don't interact with groups that disagree with them;
  • tight crowds, which are close communities, such as conference attendees, who strongly connect to each other;
  • broadcast networks, which are triggered by news media outlets and have many followers who rarely interact with each other;
  • support networks, which are headed by companies, government agencies, or organizations that respond to complaints or questions;
  • fragmented brand clusters, which are many small groups that form around products, brands, and celebrities and attract a large number of followers but foster little interconnectivity; and
  • community clusters, which are medium-sized groups that surround popular, relevant topics and develop more connectivity than their fragmented counterparts.

The last two categories are particularly relevant to marketers, with the former encompassing most marketer-customer conversations, and the latter representing where they should be falling.

"In fragmented brand clusters, marketers do most of the sharing," Smith says. "Representing their brands, they tweet out company information, product promotions, links, photos, and other content. Their followers tend to favorite or retweet the content, but rarely reply or engage in conversations. This ends the discussion quickly, resulting in only a small group formation. From a relationship-building perspective, this isn't a meaningful interaction," he says.

Instead, brands should strive to form community cluster conversations. According to Smith, when marketers and brand managers construct their messages to better relate to current events or evergreen topics important to their audiences, the resulting responses create more dialogue and interconnectivity between followers, form larger groups, and ensure better brand experiences. "Users that engage in longer conversations with their favorite brands and with others that share that brand interest are generally more invested and passionate about the brand, and that's what marketers want," Smith says.

To breed community clusters, Smith says marketers should ask more questions and encourage customers to engage, not just push out brand content. "It might feel counterintuitive not to promote new products or be proactive, but that's not what social networking is about. There's got to be a balance," he explains.

For those who are skeptical of categorizing conversations so simply, Smith suggests downloading SMRF's NodeXL and studying and visualizing social media engagements themselves.

Available as an open-source plug-in, NodeXL integrates with Excel and relies on the Clauset-Newman-Moore algorithm, a probabilistic model of hierarchical clustering for complex networks. Users just have to download the tool, import the data, and let the tool do the rest. Each resulting graph is a map of the types of real-time conversations surrounding the imported topic, which can be a hashtagged item or a Twitter handle.

The illustrations capture a snapshot of social media at any given moment, and "are immensely valuable tools for brands looking to determine how far-reaching their messaging efforts are, and ultimately, whether or not they are effective," Smith says.

The NodeXL tool can map conversations across many social channels—including foreign ones, such as microblogging site Sina Weibo in China—and track similar patterns that span cultural divides.

"It's amazing to see that despite an array of differences, people actually behave in very similar ways," says Lee Rainie, director of the Pew Research Center's Internet and American Life Project. "In the real world, crowds behave in predictable ways, and that's true on the Internet as well. Crowds have shapes that can provide key insight into human behavior, and this tool can reveal those shapes and unleash that insight."

CRM Covers
for qualified subscribers
Subscribe Now Current Issue Past Issues

Related Articles

Conversocial Launches Crowds to Facilitate Peer-to-Peer Customer Support

New platform enables customers to help each other with common support needs.

PowerObjects Releases PowerSocial, an Dynamics CRM Add-on for Social Media Management

PowerSocial Users can post and monitor Twitter and LinkedIn messages directly from Microsoft Dynamics CRM.

LeadSift Releases Self Serve Platform for SMBs

The LeadSift Self Serve platform analyzes millions of Twitter conversations in real time to find opportunities.

LogMeIn's BoldChat Adds Twitter Management

The application gives customer service teams single tool for managing customer engagement on social, mobile, email, and Web channels.