What Analytics Don’t Tell You About Social Traffic

More than half of media websites’ traffic comes from unidentified social referrals rather than social networks, The Atlantic reports.

When Chartbeat began to hone in on the unknown referrals that The Atlantic’s analytics watchers were seeing in their data, the company broke such referrals into two classes: users who came to a home page or a section landing page and those that came to a specific, and much longer link. The idea was this: People weren’t actually typing in the longer links; they were clicking on them after receiving them socially on a platform, such as instant messaging, that Chartbeat doesn’t clearly identify.

The resulting data showed that these unidentified social sources, which Madrigal dubs “dark social,” drove more users to The Altantic than all the conventional social referrers combined. At Madrigal’s request, Chartbeat produced aggregate numbers for a representative set of media sites, and the same principle applied.

Madrigal concludes:

[I]f you think optimizing your Facebook page and Tweets is “optimizing for social,” you’re only halfway (or maybe 30 percent) correct. The only real way to optimize for social spread is in the nature of the content itself. There’s no way to game email or people’s instant messages. There’s no power users you can contact. There’s no algorithms to understand. This is pure social, uncut.

But until Chartbeat starts tracking instant messaging and other “unknown” forms of link-sharing and ways to optimize for those formats emerge, there’s no alternative for publishers than to optimize for social networks as they try to bring traffic to their sites.

Related Stories
Mediabistro Course

Social Media Metrics

Social Media MetricsStarting September 4, work with a social media manager to monitor, measure and optimize your social media efforts! Danielle Brigida will teach your how to sift through web analytics, Facebook Insights, and Twitter mentions to develop a comprehensive reporting and tracking system for your brand. Register now!