In my last column I pointed out that social shares (the number of times an article is shared on social media services like Facebook and Twitter) is the universal and publicly accessible metric for judging an article, simply because it's the only universal and publicly accessible metric for judging an article.
Not exactly a controversial stance, but I suspected many would take offense at this fact. I acknowledged "Any new metric invites a vigorous debate on how healthy it is -- the value of the pageview has been hotly debated for over a decade -- and the social share has its pros and cons. I'll leave that analysis for another column."
One journalist agreed with me and tweeted "Self-promotion now annoyingly essential." Another wryly added "(Please RT :/)." Salon.com, a site that displays over a dozen social media counts on a single article page, took offense: "Apparently Galant has never heard of paychecks, jobs, party invitations, speaking gigs, prizes, television appearances and book deals to mention only a few of the earthly, though rapidly vanishing, delights of a media career."
Of course, none of those metrics are public and quantifiable. If paychecks were public, they'd certainly be paid attention to. When freelancer Noah Davis revealed in The Awl what publications paid him for his work his article was shared over 3,000 times, which included shares from about a hundred journalists. The Awl paid him $250 for the piece. While pay data would certainly be a popular metric if it were public, it probably wouldn't be a very good metric. Often the pay is higher for less prestigious journalism work, and in general compensation reflects past performance rather than current achievements.
All this leads us back to the question at hand: Are social shares a good metric to use to judge journalism?
I asked a number of savvy journalists and social media editors what they thought of social shares, which helped define the issues. Let's start by examining the biggest objections to and weaknesses of the social share:
People may share a link without reading it, or read without sharing. We all care how we're perceived on social media. Some people share articles from publications like The Economist and Harvard Business Review just to look smart, or from The Onion to appear to have a sense of humor, without actually reading them. As Amy Vernon commented on my last post, "with pageviews, you actually know someone looked at the page, for at least a brief moment." Fair point.
Social shares are just a popularity contest. Similar to pageviews, unique visitors, circulation numbers, book sales or democratic elections, social shares to some degree measure how popular something is. This objection seems to be rather reactionary. We already use plenty of metrics related to popularity to judge the impact of journalism and even to choose leaders for our nation. Part of the mission of journalism, as opposed to academic research, is to inform the public. What journalist would be satisfied if no one read their work or cared to talk about it?
Social shares are centrally controlled by corporations. While pageviews are a measure of the open web that can be tracked through a variety of means, social shares are centrally measured and reported by the social networks themselves. There may be an incentive for the networks to count these numbers liberally to appear more popular, though publishers can check if it's actually resulting in traffic. This is certainly a concern, and a subset of a larger worry many have about the web being more centralized in general.
The definition of a social share is influx. Related to the prior criticism, what goes into the social share metric can change at the whim of a social network. There are judgment calls that go into calculating what a social share is, such as counting retweets of a tweet with a link as a share itself, we might not agree with. Of course, there are plenty of judgement calls that go into defining a pageview or unique visitor by services like Google Analytics (goog) or Omniture (adbe), so this is nothing new. But the concern is heightened since each social network is the exclusive provider of its own share counts.
Autosharing increases social share counts. There are some Twitter accounts set up to tweet anything in a given RSS feed, so some articles will get a handful of tweets no matter what. When this happens it definitely skews the numbers, but since tweets are public (tweets from private Twitter accounts are not counted) it can be detected when this occurs.
Now that we've got the objections on the table, let's look at the virtues of social shares:
Social shares are an intentional action by users. It's easy to get pulled into viewing an article you don't really want to read by a misleading headline, preview photo or search result, which results in a pageview and often revenue for the publisher. But as Buzzfeed's VP Product Chris Johanesen wrote "you can’t trick someone into sharing a story with their friends." Sharing an article on social media is a very intentional act by a user, requiring several more clicks than a pageview.
Social shares require users to put their identities on the line. Unlike pageviews and comments, when people care to share an article they're willing to tell the world on Twitter or at least their friends on Facebook (fb) that they care about the article.
Social share counts don't rely on sampling. One of the few objective measures of attention on the web today is sampled data from the likes of Nielsen (nlsn) and comScore (scor). When done well, sampling can be a powerful indicator, but as many people who've miscalled elections know it's not perfect. Often minorities and new platforms are underrepresented. Sampled data can only be reported at a publication level rather than for a given article due to lack of data. Social share numbers report a full count for every article on the web.
Social share counts are public and therefore not easy to fib about. Unlike just about every other web metric, anyone can look up social share counts. Many publishers and entrepreneurs are known to manipulate the metrics they release to the public. They've been able to get away it with since only they have access to the analytics system. Social shares are out of the control of the publishers and a uniform metric.
Social shares are counted the same way on any website. Determining even seemingly simple metrics like pageviews and unique visitors is a challenging engineering problem, and different analytics packages use different techniques. Google Analytics alone uses two different techniques depending on which report you're viewing. Even if all publishers released the full data of how much traffic all of their articles are receiving, it'd vary based on which analytics tool they used. For all of social shares' shortcomings, at least they're calculated the same way on every website.
Social shares are indicative of other non-public metrics Notwithstanding the people sharing articles they've never read to look smart, we can get a rough estimate of how many people read an article based on the number of social shares. Because the number of pageviews is usually not public, social shares are the best estimate an outsider has. ShareThis, a sharing widget installed on more than two million publisher sites, has calculated that one share equals 26 clickbacks on average.
Just like any metric, the social share count has a large share of flaws. Despite that, the social share deserve serious attention for all the potential insights listed above. Most everyone I talked to about this metric said it should only be considered along with a host of other metrics. As one put it, "I tend to fall on the side of data, data, and more data -- mashing it all together for a true picture." I agree. However, the only data publicly available on an article level is social share data.
Some publishers and platforms are starting to change this. Many publishers are now allowing their journalists to access to analytic systems like Chartbeat and Clicky that allow them to see a variety of metrics on their stories in real time. Medium allows writers to compare how many "views", "reads," and "recommendations" their own article has, even computing a "read ratio." Gawker and Business Insider display view counts of each article publicly, and Gawker even shows how many new visitors the article attracted to the site. These are all great examples of metrics becoming more accessible and nuanced, but none of them are universal.
For the time being the social share is here. It's everywhere. It's got many virtues. It's got no shortage of flaws. It's unrivaled and impossible to ignore. Now please share this article with your friends.
Gregory Galant's the CEO of Muck Rack, the social network for journalists and companies in the news. He's also the cocreator of the Shorty Awards which honors the best of social media. Galant advises several startups and is a mentor in the TechStars startup accelerator. Follow him on Twitter.