Lately lots of people have been emailing me links to articles about a phenomena called "dark social." This is interesting because "dark social" is the phrase coined to describe what was originally thought to be impact of the social sharing of links directly between individuals through email and chat. Back in 2012, Alexis Madrigal wrote an article about a concerning trend that he was seeing on theatlantic.com. He observed a large number of sessions landing on deep links without any referrer data, but felt that they had to be coming from somewhere.
Let's pause for a moment for a short quiz. Before we can understand Dark Social it's important to understand what "Direct" website traffic is. Which one of the following describes what web analytics tools call "direct"?
A. Traffic that comes directly to your site via a URL that is either typed or bookmarked.
B. Traffic that comes to your site without a ‘document.referrer’
C. Improperly tagged links from email campaigns.
D. URL’s that are shared directly from one person to another via chat or email.
E. Pledge traffic responding to a on-air promo.
You're going to get the points if you said that it can actually be any, or all of those.
Measuring the amount of direct traffic to our site can help us to better understand our relationship with users and the strength of our brand. Users who come to our site via typed or bookmarked URL’s either identify the station as an important resource or have a habitual relationship with the station. However, if we take this at face value we might be missing something.
In Madrigal's more recent article he points to a new influence in why dark social is so hard to track. The social sites and apps themselves are creating confusion by failing to pass referrer data in some cases. He sums it up nicely by saying; “The takeaway is this: if you’re a media company, you are almost certainly underestimating your Facebook traffic. The only question is how much Facebook traffic you’re not counting.”
The article walks through the stunning correlation that Madrigal found when he compared traffic from facebook mobile to "dark social" traffic to articles. He explained "When mobile Facebook referrals went up or down, the dark social traffic generally moved in accordance, too." I looked at the same data for stations sites, didn't see the same tightly coupled movement, but there was a more general trend.
Over the last 13 month public radio station sites have seen increases in the amount of traffic that lands directly on article pages with no apparent source, however that traffic as a percentage of our overall traffic is flat to the prior year at about 18%. NPR.org sees even less of the dark social influence at only 8% of overall traffic. One key here is that NPR does a great job of adding campaign tracking to their URLs.
If you'd like to better understand how dark social is impacting your site there are a couple of steps you can take to better measure it. First, try looking at it in Google Analytics using a custom segment that detects sessions that land on an article page without referrer data. You should be able to use this custom segment, but double check that it matches the URL format for your pages.
Chartbeat has also done some work to troubleshoot this issue and they have published their findings and used them to update their own reporting to include what they call "direct social" traffic. Through the use of user agent data they are now able to include an additional 10-50% that was previously misattributed. They found "The most interesting observation here is how well correlated dark social traffic is to the identifiable sources. In this example, you could be convinced that the dark social is really just misattributed traffic from facebook and reddit." If you're using Chartbeat you can now find direct social traffic in multiple dashboards.
With all this fuss you might ask “why does tracking dark social matter?” Even though it’s a bit more art than science, it helps us to better understand our reliance on Facebook and other social sites to provide traffic. Paying attention to the true source of traffic prevents us from over-estimating user loyalty and brand strength. It also reminds us to keep an eye out for other potential sources of misattributed traffic.
Our best practice recommendation is to use campaign tracking in all URL’s that you are promoting either in social or email. Despite the potential for some misattribution of secondhand sharing; it can help to improve accuracy by giving you some information about how the user got to you. Google’s URL builder makes this easy to do on a case by case basis but there are also tools that will help you to track links with more consistency.
For another look feel free to watch the video from our December Station Analytics Insights Webinar on nprstations.org