I recently came across an intriguingly thoughtful article entitled “Dark Social: We Have the Whole History of the Web Wrong” on the Atlantic. It may stimulate some critical reflections regarding how researchers should study the implications of information and communication technologies (ICTs) may influence the way people share and connect with others.
The main idea of the article is simple: the recent social media phenomenon only represents a small and observable portion of the social sharing behaviors online. This is just right on the money. A good friend of mine emailed me this Dark Social article because he doesn’t really use Facebook. If I want to share this article with friends who live in China and Taiwan, I would have to email it to them, post the link on Chinese-written forums that they frequent, or send a message via the IM clients that they use. Not all my sharing practices contain referrer data and they will be in the big myth of the Dark Social. If the sharing traffic recorded by social network sites such as Facebook, YouTube, Twitter, etc. can only account for less than half of the volume, then how can researchers study the social sharing behaviors not captured by the data architecture of the so-called Web 2.0 services?
One possible starting point for answering this question is to consider people and the social contexts where these ICT-mediated practices are situated. People layer various ICTs and use their capabilities to connect with their existing personal networks without adopting the structure of the current social network sites. So, it may be more straightforward for researchers to study my anecdote by asking me than obtaining the data via various sources of online analytics behind the scene.
However, informants are nowhere near perfect. Often times, people may forget what they did and may not be able to articulate and tell researchers what they did upon inquiry. Additionally, people may not commit to provide the most truthful responses to researchers’ questioning. The challenges of Dark Social actually then become: how can researchers collect robust information beyond the data architecture of ICTs? More importantly, how can researchers benefit from the Big Data potential even if they are not necessarily interested in any particular Web 2.0 sites?
My response to these questions is going back to the basics of research design. The idea is to incorporate ICTs’ capacities into existing methodologies and research designs to help research participants provide their information to researchers. For example, the Roxy, developed by Ericka Menchen-Trevino, seeks to collect participants’ online behaviors in detail and in turn allow participants to review their actual Web usage while being interviewed by researchers. For my EgoGalaxy, I incorporated many design suggestions proposed by the literature in survey methodology to help participants recall and respond to the personal network questions. These attempts are important research design practices and potential tools for investigating the Dark Social. With the aids of technologies, people are more likely to provide robust data about their social sharing online. More importantly, researchers can collect information from the people who do not use those popular social network sites for sharing or those who purposefully avoid digital traces when they share.
Ultimately, I believe thinking about how to integrate ICTs in the existing methods will help researchers to put together a good mixed-method design for their work. The Dark Social is only invisible to some methods, and researchers can and will be able to shed light into the iceberg.
Comments are closed.
Couple things: 1. This guy seems to put too much emphasis on sharing links as the purpose of the social web. From his perspective, it’s the most important thing because he wants traffic. And I get that, I used to be that person for a media company too. But, sharing links is only 1 facet of social media use for many users. So, when we study Twitter, FB we’re not just studying information flows. If that’s all we did, everyone would be SNA people and we’d call it a day. Clearly, there’s more to it than that. 2. I hate to rain on his parade, but using your example Patrick, let’s flip it and say that you first saw the article because I posted it on FB and THEN you emailed/IM’d it to your friends. The point is, I would bet that if you took all the social sharing features off all the stories on the Atlantic (made it harder for people to share on FB and Twitter, etc.) that a significant drop in “Dark Social” would occur as a result. His error here is that in thinking because there is a metric for particular sites, it’s the only measure of their influence. He’s dead wrong.
In my original example, none of the shares were captured, which will be what happens if there is no social media sharing features on the Atlantic. In your revised example, the part that you shared with me is captured when I clicked the link that you posted, the latter part of emailing and IMing is not. So, if you took all the social sharing features off all the stories on the Atlantic (made it harder for people to share on FB and Twitter, etc.), then I think the Dark Social will become LARGER because we lost the referrer metadata and thus we have less information about the number of shares of these articles.
Hey Patrick, what I mean is one post to a social site can generate many cut and pastes of the link. This is hard to explain but his idea that because none of the Dark social click-throughs can be directly accounted for means that social media had NO role in their occurring. I think that’s wrong. My point, to put it another way, is that If there were no social media shares of the site, overall site traffic would drop significantly (not just the X% of clicks that FB gets but also an X proportion of the Dark social). I think he underestimates the role of social sites in the distribution of individual page links. So yes, by necessity the Dark social clicks would increase but overall site traffic would decrease. In a big way.
I see your point and I would suspect that you are right about the total traffic, Stacy . However, we need to think how more difficult would it be to copy the link and paste somewhere to share with your friends than clicking the FB share button on sites. We also need to think whether the benefits of keeping your privacy from social media may outweigh such an additional cost. I would suspect that much of the Dark Social traffic may be the result of the privacy concern. If this is really the case, then it means that people share things even more actively then we know, and the overall traffic may not necessarily drop much even without social media.